The deployment of self-contained machine learning technology running on high-performance yet low-power devices is a difficult challenge. The inevitable slowing of Moore's law also suggests technology scaling alone will not provide a simple solution. Furthermore, highly specialized accelerators, while efficient, may lack the flexibility required to support a broad range of evolving techniques or the possibility to make best use of the available resources. In this talk I will outline some of the work we have done in this area and suggest avenues for future research.

Dr. Robert Mullins is a Senior Lecturer in the Computer Laboratory at the University of Cambridge. His research and teaching focuses on computer architecture and VLSI (chip) design. He is a Fellow of St. John's College and Director of Studies for Computer Science at St. John's College and Peterhouse College. He was a founder of the Raspberry Pi Foundation, a UK charity that promotes the study of computer science and electronics at the school level. He is a founder and director of the lowRISC project.